In their drive to maximize revenue streams, modern broadcasters are continuously exploring more services for linear and on-demand content delivery. From 3-D hybrid TVs to smartphones, they all offer viewers numerous options for accessing content while selecting languages, audio and video quality, subtitling, background information, interaction, etc. More devices that offer even more user options have not only increased the complexity of digital rights management, but also the number of files that is required in a multitude of formats (to air only one piece of content) is heading for new heights. In this multiplatform environment, verifying whether the rights are available for all desired services is no longer a simple “yes” or “no” query. In ever more cases, complex and expensive rights dictate that not all content can be made available on every platform. And if content is not delivered on a particular platform, it is undesirable that transcoding capacity and storage space would be wasted for file formats that will never be used. In practice, this means that for every title a broadcaster needs to deliver, the rights, services and marketing will dictate which files need to be available in which formats. As a result, creation of multiplatform and ready-to-air content becomes increasingly complex. It requires both expensive human resources to manage the rights, as well as expensive transcoding and storage capacity to manipulate the files.
From traditional to integrated workflow
Today, broadcasters that are still working with traditional isolated systems have become exceptions. Integrated channel management has replaced their existing systems, which were often developed in-house. Such station-wide systems manage a lot of the back office metadata and workflow, but also integrate this information on a system level with finance, administration and the broadcast environment.
For example, material status and transcoding requests flow automatically from scheduling to their MAM systems, and costs and amortization details automatically update finance, while all listings and EPG information is generated automatically. Such a fully integrated environment has long been the “Valhalla” that vendors have strived for.
With the rise of the ever more complex rights and the proliferation of multimedia devices, new challenges lay ahead, while the file jungle grows denser. Take the following example: In the old non-integrated workflow, a trailer producer received a request to create a trailer during a meeting or via e-mail. The trailer producer then retrieved the tape from the archive, edited the content, copied the trailer with graphics back to tape, copied the trailer on a new tape, and returned that tape to the tape manager. The tape manager then checked the material quality and registered the time codes and status so that it could be scheduled and aired. During the broadcast day, the voice-over was recorded, or the live announcer informed the viewer what was about to happen.
Today, we almost forget how schedule changes or misfiled tapes caused numerous errors and problems. In our current integrated environment, such mistakes are prevented. Changes in titles or starting times automatically update graphics engines, which will render the new and updated message on-screen. Tape numbers do not need to be re-keyed, and file IDs are automatically communicated from the editing suite to the MAM and then to the scheduling environment and playout suite. This improved efficiency was — and still is — an enabler for modern broadcasters, allowing them to experiment with new services, launch new channels in a swift and efficient way, and minimize the financial impact on the organization by keeping the increase in headcount to a minimum.
A new era
Let's revisit our example. In current integrated channel management systems, trailer production requests are generated from the scheduling system, which creates and maintains the metadata that describes the trailer and its campaign automatically. Requesting a new trailer automatically creates material numbers, ingest requests, transcoding tasks and subtitle requests, and it indicates that trailer clips need to be produced. This entire sequence is generated only from a simple trailer request function from the schedule, sometimes even months before the actual airing of the program and the trailer. Multiply the number of requests over the whole bouquet of channels, and you get hundreds of open trailer requests that all need to be managed by a certain day. This illustrates the weakness of the current integrated system workflow: Everybody expects that everybody knows what has to be done simply by actively looking for new content and material states in the integrated software environment.
The reality, however, is that people do not like to continually check the status of all relevant objects. Sometimes tasks are not clearly assigned to an operator and, hence, they are never properly checked. Often the point at which somebody notices items with a problem is just when the content is needed. By that time, it is usually too late to solve the problem. If a trailer scheduler notices, for example, a trailer that is not ready to air, it is unlikely that the artwork, QC process, subtitling and the voice-over processes can be finished on time.
So even in a fully integrated and modern environment, it is sometimes hard to find out what the actual status is. Maybe the process has already started, but somebody just forgot to update the system. Or, that person will only update the state when the entire job is finished, in which case it is impossible to manage the work in progress.
Workflow management — the broadcast way
The latest channel management systems come with integrated workflow functionality, designed to solve and manage workflows in a modern broadcast environment. In order to solve workflow issues, the administrators can now set up workflow scenarios that define the sequence of tasks that are required to successfully complete a particular job, such as the trailer example above.
The sequences of tasks in a workflow are defined in a particular order. This order implies that a particular task can only start when one or more preceding tasks have been completed. Only tasks for which all prerequisite tasks are actually completed will be presented to the users in a to-do list.
For example, the step that defines that the voice-over needs to be recorded is only set as to-do if the video production is finished. If a task is on a to-do list, it will wait for the acknowledgement of a user to indicate that a task is complete. If a manager must acknowledge that a particular production can start, then the workflow will contain a task called “acknowledge production.” This task will wait for the indication of a user (the manager in this case) that the acknowledgement for the production in question has been given. In other words, the acknowledge production task will be indicated as complete.
Some tasks can also be automated, meaning that the workflow task will not be marked as complete by a user, but that the task will rely on a small software routine called an actuator. Actuators perform a specific test and — based on the result of the test — the task can be put to complete. A good example of an automated task is the test if a file has been ingested or not. The ingest server processes files based on its own priority and capacity. But when a particular file is ingested, a status is set within the MAM system. The actuator in the workflow tool will test for the ingest file status by use of an API call or web service. If the actuator detects that the file is ingested, the corresponding workflow step is automatically marked as complete. Such an automated task can, hence, mark a step in the workflow as complete, which will update the to-do status of the next tasks in the workflow chain.
Automated tasks are important components that can test for information in the scheduling system itself, but also in integrated systems, such as file servers, MAM, transmission suites, file servers etc. Actuators can even be used to initiate tasks, such as file transfer via FTP, sending mails, etc. An important evaluation in anybody's to-do list is the priority. If many tasks are on your queue, it is important to be able to decide what to do first. During the creation of a workflow scenario, the administrator can indicate the expected deadlines. A deadline can be set with different rules. Deadline rules can be a simple offset in relation to the start time of the job — for example, “Create a subtitle file. The deadline is within four days from now.” A deadline can also be defined in relation to a previous task. For example: “Create a subtitle file. The deadline is within four days after the video material has been ingested.” Or a deadline can be in relation to business data from the scheduling system or any other integrated system — for example, “Create a subtitle file. The deadline is five days before the first air date.” Mail notification is a powerful tool that helps drive workflow. But as with any set-up of computer generated mails, it is imperative to maximize its efficiency. E-mails have to be sent to the correct person with information that is relevant, while preventing the sending of mails on every status change or event. Otherwise, e-mail notifications will quickly be considered as spam and will be counterproductive. The workflow administrator should be able to tune who will receive a mail notification (an individual user, a group of people, the manager, etc.) for every state or urgency level (urgent, late, very late, failed, etc.).
A typical workflow server logs all status changes and notifications that were raised. This is invaluable data to identify bottlenecks, weaknesses and opportunities in the workflow. Such data will help to tune the workflow and to let managers assign resources where they are most needed. It can even help them to estimate the impact on the staff and infrastructure when they want to launch new services.
The integrated workflow management system will help broadcasters go beyond the promise of integrated channel management systems at a fraction of the cost of traditional stand-alone products. Users not only will have access to the correct and relevant data to do their jobs, but also the system will now indicate which tasks need their attention, provide a comprehensive overview of the tasks that need to be performed, and alert which tasks are late or not yet assigned to a person. Updating and notifying the relevant person in the organization when deadlines are near will prevent mistakes and, more importantly, enable users and managers to detect problems before it is too late to take the appropriate action.
Michel Beke is product manager and co-owner of MediaGeniX.